Miss any of our Open RFC calls?Watch the recordings here! »

@modelx/model

1.6.1 • Public • Published

@modelx/Model - Machine Learning and Neural Networks with Tensorflow

Coverage Status Build, Test & Coverage

Getting started

Clone the repo and drop your module in the src directory.

# Install Prerequisites 
$ npm install rollup typedoc jest sitedown --g

Basic Usage

$ npm run build #builds type declarations, created bundled artifacts with rollup and generates documenation 

Introduction

This library is a compilation of model building modules with a consistent API for quickly implementing Tensorflow at edge(browser) or any JavaScript environment (Node JS / GPU).

Read the manual

List of Tensorflow models

Classification

Regression

Artificial neural networks (ANN)

LSTM Time Series

Basic Usage

TensorScript is and ECMA Script module designed to be used in an ES2015+ environment, if you need compiled modules for older versions of node use the compiled modules in the bundle folder.

Please read more on tensorflow configuration options, specifying epochs, and using custom layers in configuration.

Regression Examples

import { MultipleLinearRegression, DeepLearningRegression, } from '@modelx/model';
import ms from 'modelscript';
 
async function main(){
  const independentVariables = [ 'sqft', 'bedrooms',];
  const dependentVariables = [ 'price', ];
  const housingdataCSV = await ms.csv.loadCSV('./test/mock/data/portland_housing_data.csv');
  const DataSet = new ms.DataSet(housingdataCSV);
  const x_matrix = DataSet.columnMatrix(independentVariables);
  const y_matrix = DataSet.columnMatrix(dependentVariables);
  const MLR = new MultipleLinearRegression();
  await MLR.train(x_matrix, y_matrix);
  const DLR = new DeepLearningRegression();
  await DLR.train(x_matrix, y_matrix);
  //1600 sqft, 3 bedrooms
  await MLR.predict([1650,3]); //=>[293081.46]
  await DLR.predict([1650,3]); //=>[293081.46]
}
main();

Classification Examples

import { DeepLearningClassification, } from '@modelx/model';
import ms from 'modelscript';
 
async function main(){
  const independentVariables = [
    'sepal_length_cm',
    'sepal_width_cm',
    'petal_length_cm',
    'petal_width_cm',
  ];
  const dependentVariables = [
    'plant_Iris-setosa',
    'plant_Iris-versicolor',
    'plant_Iris-virginica',
  ];
  const housingdataCSV = await ms.csv.loadCSV('./test/mock/data/iris_data.csv');
  const DataSet = new ms.DataSet(housingdataCSV).fitColumns({ columns: {plant:'onehot'}, });
  const x_matrix = DataSet.columnMatrix(independentVariables);
  const y_matrix = DataSet.columnMatrix(dependentVariables);
  const nnClassification = new DeepLearningClassification();
  await nnClassification.train(x_matrix, y_matrix);
  const input_x = [
    [5.1, 3.5, 1.4, 0.2, ],
    [6.3, 3.3, 6.0, 2.5, ],
    [5.6, 3.0, 4.5, 1.5, ],
    [5.0, 3.2, 1.2, 0.2, ],
    [4.5, 2.3, 1.3, 0.3, ],
  ];
  const predictions = await nnClassification.predict(input_x); 
  const answers = await nnClassification.predict(input_x, { probability:false, });
  /*
    predictions = [
      [ 0.989512026309967, 0.010471616871654987, 0.00001649192017794121, ],
      [ 0.0000016141033256644732, 0.054614484310150146, 0.9453839063644409, ],
      [ 0.001930746017023921, 0.6456733345985413, 0.3523959517478943, ],
      [ 0.9875779747962952, 0.01239941269159317, 0.00002274810685776174, ],
      [ 0.9545140862464905, 0.04520365223288536, 0.0002823179238475859, ],
    ];
    answers = [
      [ 1, 0, 0, ], //setosa
      [ 0, 0, 1, ], //virginica
      [ 0, 1, 0, ], //versicolor
      [ 1, 0, 0, ], //setosa
      [ 1, 0, 0, ], //setosa
    ];
   */
}
main();
import { LogisticRegression, } from '@modelx/model';
import ms from 'modelscript';
 
async function main(){
  const independentVariables = [
    'Age',
    'EstimatedSalary',
  ];
  const dependentVariables = [
    'Purchased',
  ];
  const housingdataCSV = await ms.csv.loadCSV('./test/mock/data/social_network_ads.csv');
  const DataSet = new ms.DataSet(housingdataCSV).fitColumns({ columns: {Age:['scale','standard'],
  EstimatedSalary:['scale','standard'],}, });
  const x_matrix = DataSet.columnMatrix(independentVariables);
  const y_matrix = DataSet.columnMatrix(dependentVariables);
  const LR = new LogisticRegression();
  await LR.train(x_matrix, y_matrix);
  const input_x = [
    [-0.062482849427819266, 0.30083326827486173,], //0
    [0.7960601198093905, -1.1069168538010206,], //1
    [0.7960601198093905, 0.12486450301537644,], //0
    [0.4144854668150751, -0.49102617539282206,], //0
    [0.3190918035664962, 0.5061301610775946,], //1
  ];
  const predictions = await LR.predict(input_x); // => [ [ 0 ], [ 0 ], [ 1 ], [ 0 ], [ 1 ] ];
}
main();

Time Series Example

import { LSTMTimeSeries, } from '@modelx/model';
import ms from 'modelscript';
 
async function main(){
  const dependentVariables = [
    'Passengers',
  ];
  const airlineCSV = await ms.csv.loadCSV('./test/mock/data/airline-sales.csv');
  const DataSet = new ms.DataSet(airlineCSV);
  const x_matrix = DataSet.columnMatrix(independentVariables);
  const TS = new LSTMTimeSeries();
  await TS.train(x_matrix);
  const forecastData = TS.getTimeseriesDataSet([ [100 ], [200], [300], ])
  await TS.predict(forecastData.x_matrix); //=>[200,300,400]
}
main();

### Special Thanks

License

MIT

Install

npm i @modelx/model

DownloadsWeekly Downloads

49

Version

1.6.1

License

none

Unpacked Size

24.2 MB

Total Files

153

Last publish

Collaborators

  • avatar